PUBLIC HUMAN ASSAULT PREDICTION USING
HUMAN ACTIVITY RECOGNTION
EMBEDDED SYSYTEM WITH MACHINE LEARNING
Presented by,
K HARI HARAN (312421106045)
B HARI PRASATH (312421106051)
Guided by,
Dr. S TEPHILLAH (M.E.,PhD)
ASSISSTANT PROFESSOR
OBJECTIVE
• To develop a highly accurate Human Activity recognition
model.
• To develop a efficient model for real time deployment.
• To make this model as surveillance solutions for road
surveillance.
• To develop a mobile application for providing alert to alert to
the Police in case the theft detection by detecting the human
activities.
LITERATURE SURVEY
3
4
S.NO. TITLE OF
THE PAPER
WITH
AUTHOR
NAME
JOURNAL
NAME
YEAR
OF
PUBLIC
ATION
METHODOLOGY PROS &
CONS
5 An Integrated
Cloud-Based
Smart Home
Management
System with
Community
Hierarchy
IEEE
2016
6 Presentation
Attack Detection
for Face
Recognition using
Light Field
Camera
IEEE
2015
5
S.NO. TITLE OF THE
PAPER WITH
AUTHOR NAME
JOURNAL
NAME
YEAR OF
PUBLICA
TION
METHODOLOGY PROS &
CONS
7
8
6
EXISTING SYSTEM
• The proposed system uses Dual Stacked Autoencoders for Feature
Embedded Clustering (DSAFEC) and a BOW construction method based
on DSAFEC (B-DSAFEC) to improve human activity recognition from
videos.
• DSAFEC transforms video feature points into a learned feature space and
predicts their cluster assignment probabilities.
• B-DSAFEC uses these probabilities to build Bags of Words (BOWs).
• Soft clustering is applied by assigning each feature point to multiple
clusters based on the highest probabilities, reducing computational
complexity and eliminating selection restrictions.
08/02/24 7
EXISTING SYSTEM vs PROPOSED
SYSTEM
Parameters Existing system Proposed system
1.Delay
2.Dat rate
3.Energy
100ms 50ms
08/02/24 8
PROPOSED SYSTEM
• The project involves using the KTH video dataset to design a system for
human activity recognition in road surveillance.
• Preprocessing extracts specific frames, and features are obtained using
Pixel and Optical flow techniques.
• Data visualization is used to display these features.
• A Spatio-Temporal Net deep learning algorithm classifies human
activities.
• If abnormal behavior like fights is detected, an alert is sent to the police.
• A mobile app developed with React Native notifies the police and
enables live streaming.
• This system aims to prevent and provide evidence of abnormal activities
on roads in real-time.
08/02/24 9
BLOCK DIAGRAM OF THE
PROPOSED SYSTEM
08/02/24 10
BLOCK DIAGRAM DESCRIPTION
• 1. Video Input (KTH Dataset)
• Purpose: Source of video data for the system.
• 2. Preprocessing
• Purpose: Extracts specific frames for analysis.
• 3. Feature Extraction
• Purpose: Obtains features using Pixel and Optical Flow techniques.
• 4. Spatio-Temporal Net (Deep Learning Model)
• Purpose: Classifies human activities from extracted features.
• 5. Abnormal Activity Detection
• Purpose: Identifies abnormal behaviors like fights.
• 6. Alert System
• Purpose: Sends alerts to the police.
• 7. Mobile Application (React Native)
• Purpose: Notifies police and enables live streaming.
08/02/24 11
CIRCUIT DIAGRAM
08/02/24 12
ALGORITHM
08/02/24 13
EXPECTED OUTPUT
08/02/24 14
RESULTS
• Simulation output (or) Graphs or results so far
completed .
(depends on the project)
08/02/24 15
CONCLUSION & FUTURE
ENHANCEMENT
08/02/24 16
REFERENCES
[1] ature Embedded Learning for Human Activity Recognition Ting Wang, Student
Member, IEEE, Wing W. Y. Ng*, Senior Member, IEEE, Jinde Li, Qiuxia Wu, Member,
IEEE, Shuai Zhang, Member, IEEE, Chris Nugent, Senior Member, IEEE, Colin Shewell,
Member, IEEE[2021]
[2] PrivHome: Privacy-Preserving Authenticated Communication in Smart Home
Environment, Geong Sen Poh, Prosanta Gope, Member, IEEE , and Jianting Ning [2019,
VOL. 99, NO. 99]
[3] Continuous Human Motion Recognition With a Dynamic Range-Doppler Trajectory
Method Based on FMCW Radar, Chuanwei Ding , Student Member, IEEE, Hong Hong ,
Member, IEEE, Yu Zou, Student Member, IEEE, Hui Chu , Xiaohua Zhu, Member, IEEE,
Francesco Fioranelli , Member, IEEE, Julien Le Kernec , Senior Member, IEEE, and
Changzhi Li , Senior Member, IEEE [2019, VOL. 57, NO. 9, ]
17
08/02/24
REFERENCES
[4] PRNU-Based Camera Attribution from Multiple Seam-Carved Images,
BSamet Taspinar, Manoranjan Mohanty, and Nasir Memon.
[2017, VOL. 20, NO. 5, ]
[5] An Integrated Cloud-Based Smart Home Management System with
Community Hierarchy, Ying-Tsung Lee, Wei-Hsuan Hsiao, Chin-Meng Huang
and Seng-Cho T. Chou. [2016, Vol No: 2162-237X]
[6] Presentation Attack Detection for Face Recognition using Light Field
Camera, R. Raghavendra Kiran B. Raja Christoph Busch Norwegian Biometric
Laboratory, Gjøvik University College, Norway. [TIP.2015.2395951]
18
08/02/24

Template_project_review_1_Phase_I[1][1].ppt

  • 1.
    PUBLIC HUMAN ASSAULTPREDICTION USING HUMAN ACTIVITY RECOGNTION EMBEDDED SYSYTEM WITH MACHINE LEARNING Presented by, K HARI HARAN (312421106045) B HARI PRASATH (312421106051) Guided by, Dr. S TEPHILLAH (M.E.,PhD) ASSISSTANT PROFESSOR
  • 2.
    OBJECTIVE • To developa highly accurate Human Activity recognition model. • To develop a efficient model for real time deployment. • To make this model as surveillance solutions for road surveillance. • To develop a mobile application for providing alert to alert to the Police in case the theft detection by detecting the human activities.
  • 3.
  • 4.
  • 5.
    S.NO. TITLE OF THEPAPER WITH AUTHOR NAME JOURNAL NAME YEAR OF PUBLIC ATION METHODOLOGY PROS & CONS 5 An Integrated Cloud-Based Smart Home Management System with Community Hierarchy IEEE 2016 6 Presentation Attack Detection for Face Recognition using Light Field Camera IEEE 2015 5
  • 6.
    S.NO. TITLE OFTHE PAPER WITH AUTHOR NAME JOURNAL NAME YEAR OF PUBLICA TION METHODOLOGY PROS & CONS 7 8 6
  • 7.
    EXISTING SYSTEM • Theproposed system uses Dual Stacked Autoencoders for Feature Embedded Clustering (DSAFEC) and a BOW construction method based on DSAFEC (B-DSAFEC) to improve human activity recognition from videos. • DSAFEC transforms video feature points into a learned feature space and predicts their cluster assignment probabilities. • B-DSAFEC uses these probabilities to build Bags of Words (BOWs). • Soft clustering is applied by assigning each feature point to multiple clusters based on the highest probabilities, reducing computational complexity and eliminating selection restrictions. 08/02/24 7
  • 8.
    EXISTING SYSTEM vsPROPOSED SYSTEM Parameters Existing system Proposed system 1.Delay 2.Dat rate 3.Energy 100ms 50ms 08/02/24 8
  • 9.
    PROPOSED SYSTEM • Theproject involves using the KTH video dataset to design a system for human activity recognition in road surveillance. • Preprocessing extracts specific frames, and features are obtained using Pixel and Optical flow techniques. • Data visualization is used to display these features. • A Spatio-Temporal Net deep learning algorithm classifies human activities. • If abnormal behavior like fights is detected, an alert is sent to the police. • A mobile app developed with React Native notifies the police and enables live streaming. • This system aims to prevent and provide evidence of abnormal activities on roads in real-time. 08/02/24 9
  • 10.
    BLOCK DIAGRAM OFTHE PROPOSED SYSTEM 08/02/24 10
  • 11.
    BLOCK DIAGRAM DESCRIPTION •1. Video Input (KTH Dataset) • Purpose: Source of video data for the system. • 2. Preprocessing • Purpose: Extracts specific frames for analysis. • 3. Feature Extraction • Purpose: Obtains features using Pixel and Optical Flow techniques. • 4. Spatio-Temporal Net (Deep Learning Model) • Purpose: Classifies human activities from extracted features. • 5. Abnormal Activity Detection • Purpose: Identifies abnormal behaviors like fights. • 6. Alert System • Purpose: Sends alerts to the police. • 7. Mobile Application (React Native) • Purpose: Notifies police and enables live streaming. 08/02/24 11
  • 12.
  • 13.
  • 14.
  • 15.
    RESULTS • Simulation output(or) Graphs or results so far completed . (depends on the project) 08/02/24 15
  • 16.
  • 17.
    REFERENCES [1] ature EmbeddedLearning for Human Activity Recognition Ting Wang, Student Member, IEEE, Wing W. Y. Ng*, Senior Member, IEEE, Jinde Li, Qiuxia Wu, Member, IEEE, Shuai Zhang, Member, IEEE, Chris Nugent, Senior Member, IEEE, Colin Shewell, Member, IEEE[2021] [2] PrivHome: Privacy-Preserving Authenticated Communication in Smart Home Environment, Geong Sen Poh, Prosanta Gope, Member, IEEE , and Jianting Ning [2019, VOL. 99, NO. 99] [3] Continuous Human Motion Recognition With a Dynamic Range-Doppler Trajectory Method Based on FMCW Radar, Chuanwei Ding , Student Member, IEEE, Hong Hong , Member, IEEE, Yu Zou, Student Member, IEEE, Hui Chu , Xiaohua Zhu, Member, IEEE, Francesco Fioranelli , Member, IEEE, Julien Le Kernec , Senior Member, IEEE, and Changzhi Li , Senior Member, IEEE [2019, VOL. 57, NO. 9, ] 17 08/02/24
  • 18.
    REFERENCES [4] PRNU-Based CameraAttribution from Multiple Seam-Carved Images, BSamet Taspinar, Manoranjan Mohanty, and Nasir Memon. [2017, VOL. 20, NO. 5, ] [5] An Integrated Cloud-Based Smart Home Management System with Community Hierarchy, Ying-Tsung Lee, Wei-Hsuan Hsiao, Chin-Meng Huang and Seng-Cho T. Chou. [2016, Vol No: 2162-237X] [6] Presentation Attack Detection for Face Recognition using Light Field Camera, R. Raghavendra Kiran B. Raja Christoph Busch Norwegian Biometric Laboratory, Gjøvik University College, Norway. [TIP.2015.2395951] 18 08/02/24